Appropriate mathematical and statistical models will be developed to investigate the genetic components of cancer etiology in large families: (1) models will be developed for the aging process, with a view to determining from the pattern of mortality whether the individuals in the families fall into two broad classes, those genetically disposed to a specific disease and those not so disposed; (2) statistical models will be developed for the pedigree analysis (segregation and linkage) of susceptibility to a disease with variable age of onset. Computer programs based on these models will be written and applied to data collected by others on extensive cancer families. Blood samples from these same families will be analyzed for as many polymorphic genetic markers as possible, and linkage relationships between cancer susceptibility and each of the genetic markers will be investigated. Other marker variables that may be related to cancer proneness will also be taken into account, to sharpen the genetic analysis. This will result in (1) the development of appropriate theory, based on the results obtained, for genetic counseling of individuals in cancer families, and (2) the possibility of classifying individuals with regard to their genetic susceptibility to cancer, and hence provide genetically homogeneous groups of individuals for the study of environmental causes and the biochemical pathways involved. BIBLIOGRAPHIC REFERENCE: Elandt-Johnson, R. C. "On various estimators of qi based on incomplete failure data in follow-up studies", To appear in Journal of Chronic Disease, vol. 30, 1977.